In recent years mesoscale weather prediction models have become increasing common at the National Weather Service Weather Forecast Office (WFO) level. Advances in computing power have allowed mesoscale models to be run real time on relatively inexpensive desktop computers. The result is that many weather offices tailor the parameters of these models for local use. The purpose of customized mesoscale models at the WFO level is to provide guidance for local mesoscale phenomenon that would otherwise be overlooked by regional models. The Tallahassee, Florida National Weather Forecast Office is currently evaluating two mesoscale weather prediction models, the MM5 and the Workstation ETA (WS-ETA). The WS-ETA is run by the Tallahassee WFO and can be reconfigured as desired. The MM5 is being run by the Florida Division of Forestry, and output from the model is made available to the Tallahassee WFO. Both models are run twice daily and use a 15-km grid resolution domain that includes the panhandle of Florida. We propose to evaluate the mesoscale model guidance in a real time environment through the spring and early summer seasons. Forecasters will log their twice-daily evaluation of the guidance from both the MM5 and WS-ETA. Furthermore, specific weather events will be modeled to evaluate performance based on varying parameters such as physical schemes and resolution changes. The goal of this study is to determine if 1) Mesoscale models provide value-added guidance over regional scale models, 2) Which mesoscale models provide better guidance both in general and for particular mesoscale phenomenon.